Fix CooMatrix.to_dense for duplicate indices ( #8187).Fast IndexIterator for ChainerX CUDA ( #8360).Support n_step_gru function on exporting ONNX ( #8492, thanks Extend ONNX-Chainer's TransposeSequence converter to support more cases ( #8493, thanks Add NStepGRU link converter example to ONNX-Chainer test ( #8494, thanks Allow ONNX-Chainer's patch_functions to patch functions in modules other than chainer.functions ( #8495, thanks Replaced n_fold with n_folds ( #8516, thanks Remove trailing whitespaces ( #8536).
Support custom initializers in NStepRNN ( #8489).See here for the complete list of solved issues and merged PRs. Ignore NumPy 1.20 deprecations ( #8598).Mitigate breaking changes in CuPy v8 ( #8583).Avoid pytest.PytestUnknownMarkWarning ( #8599).Use stable CuPy v7 in ONNX test ( #8592).Ignore CuPy deprecation warnings in tests ( #8589).Fix simple typo, cotiguousness -> contiguousness ( #8595, thanks Add CuPy version recommendation ( #8608).Import first to show preload warning ( #8605).Fix chainer.testing requiring pytest installed ( #8611).Support importing cuDNN in CuPy v8 ( #8590).Add CHAINER_WARN_VERSION_MISMATCH environment variable ( #8588).There are no further planned releases for Chainer v7 series. Please also understand that CuPy v10 is not compatible with Chainer.Īs announced previously, Chainer is under the maintenance phase. A warning message will be shown if you run Chainer v7 with CuPy v8 or later, but you can disable it by setting the new CHAINER_WARN_VERSION_MISMATCH=0 environment variable. We confirmed that basic tests and examples run fine, but we still encourage Chainer v7 users to stay with CuPy v7.8.0 & CUDA 10.2 or earlier & cuDNN v7.6 if you don't have strong reasons to upgrade. See here for the complete list of solved issues and merged PRs.įor those who need to run Chainer on CUDA 11.1+, this release provides "limited" support for CuPy v8/v9.
0 kommentar(er)
